Retrieval of the Diffuse Attenuation Coefficient from GOCI Images Using the 2SeaColor Model: A Case Study in the Yangtze Estuary

Xiaolong Yu, University of Massachusetts Boston, School for the Environment, Boston, MA, United States, Mhd.Suhyb Salama, Faculty of Geo-Information Science and Earth Observation, University of Twente., ITC, Enschede, Netherlands, Wouter Verhoef, University of Twente, Faculty of Geo-Information Science and Earth Observation (ITC), Enschede, Netherlands and Fang Shen, East China Normal University, State Key Laboratory of Estuarine and Coastal Research, Shanghai, China
The diffuse attenuation coefficient (Kd) is an important optical property in water body that related to penetration of solar radiance. Remotely sensed retrieval of Kd is one of the hottest topics in ocean color researches and yet still full of challenges for turbid waters. The 2SeaColor model (Salama and Verhoef, 2015) was proposed to analytically retrieve Kd from remote sensing reflectance (Rrs) for both clear and turbid water, but its performance over turbid water samples are submitted to large uncertainties. In this study, we present an improvement on the parametrization equations in the inverse scheme of the 2Seacolor model. The improved 2SeaColor model is then validated and compared with the Zhang model and the Lee model using three in-situ datasets.

Validation with radiometric data from field observations shows that the 2SeaColor model generates best estimates of Kd for the full range of observations, with the largest determination coefficient (R2 = 0.935) and the smallest root mean square error (RMSE = 0.078 m-1). Specifically, the Zhang model and the Lee model provide better estimation for clear water where Kd (490) < 0.2, while results from 2SeaColor is rather comparable. As for turbid water where Kd (490) > 0.2 m-1, 2SeaColor model is verified more appropriate with RMSE of 0.186m-1, compared to RMSE of 0.279 m-1 and 0.256 m-1 for the Zhang model and the Lee model, respectively.

The 2SeaColor model is last applied on the GOCI (Geostationary Ocean Color Imager) level 2 product (L2P) to produce Kd maps over the Yangtze estuary, resulting in a reasonable distribution and expected range of Kd as Kd (490) varying from 0.04 to 9.82 m-1 for the image acquired at 02:16 UTC, March 8 of 2013.

The analytical 2SeaColor model is able to provide consistently stable and accurate Kd estimates in both clear and turbid waters without the need of tuning empirical coefficients from field measurements, and thus has great potential for estimating Kd over optically complex waters.